The Origin of Electoral Systems in the Postwar Era

A worldwide approach

Krister Lundell

• Publisher: Routledge
• Year: 2009
• Print ISBN: 9780415477147
• eBook ISBN: 9781135230661
• Edition: 1
• Topic: Politics & International Relations
• Subtopic: International Relations
• Index: Politics & International Relations

This book explores and presents the influence of contextual factors on the choice of electoral systems for parliamentary elections in both democracies and non-democracies around the world.

Taking a macroscopic approach, the author focuses on structural explanations, with an emphasis on general patterns rather than country specific explanations. Drawn from a wealth of data, the book presents the frequency of the adoption of each electoral formula and system in the postwar era and is followed by a theoretical elaboration of electoral system choice. The author then draws on rational, cultural/historical and institutional theories which are systematically analyzed by means of sophisticated bivariate and multivariate techniques. Lundell demonstrates that few electoral systems have been chosen from rational considerations and the impact of the cultural and historical setting is tremendous; colonial legacy, regional influence and temporal trends largely explain the cross-national variation in electoral systems.

This book will be of interest to students and scholars of political science, electoral politics and comparative politics.

1 Introduction

Objectives and scope

The traditional way of studying electoral systems is to focus on the consequences of different electoral rules. Electoral system research in the twentieth century also encompasses strategic effects of electoral methods, that is, the nature and extent of choice given to voters under different systems. Since the fall of the Iron Curtain and the breakdown of Communism, a growing trend in the literature has been to understand the basis for electoral system choice rather than the consequences of electoral laws. The wave of democratization in Third World countries and the debate on electoral reform in advanced democracies since the early 1990s have also contributed to the trend of studying electoral systems as explanandum instead of explanans. The present book belongs to this field of electoral system research.

Until the late 1950s, there was a strong belief that political results were almost completely determined by the institutional framework. Maurice Duverger, one of the foremost theorists within this approach, emphasized the importance of electoral rules in explaining the nature of a country’s party system (1951). The growing influence of sociological methods and theories in political science in the 1950s called this view into question by emphasizing the role of social phenomena as the main determinants of party systems. Party Systems and Voter Alignments: Cross- national Perspectives, written by Seymour Martin Lipset and Stein Rokkan (1967) is perhaps the most central work within the sociological approach. The authors maintained that the emergence of party systems in the industrialized world is explained by four lines of social cleavage: centre/periphery, state/church, land/industry and owner/worker. Modern party systems are seen as a result of the kinds of cleavages that were present in society when political parties were organized, and how conflicts were managed. Electoral systems only modified socially determined party systems.

Although major historical events and ethnic, religious and economical cleavages cannot be neglected in the study of electoral institutions, several research efforts since the 1960s have confirmed the importance of electoral rules in shaping politics. Both approaches bear relevance to this book. The contextual environment provides the basic theoretical framework for explaining electoral system choice. However, the third wave of democratization has given rise to a similar third wave of electoral system choice, in which deliberate electoral system design has played a central role. The virtues of different electoral systems have been widely debated on the basis of research on political consequences of electoral rules. The sociological and the institutional approach are actually related to each other: the former regards the electoral system as an expression of social cleavages and party systems, whereas the latter regards the electoral system as a manipulative instrument for shaping politics in accordance with societal requirements. Thus, both represent a rational perspective on the role of electoral systems. However, it should be observed that this study is not concerned with rational choice analysis. In this context, the term rational means that the structural traits of a society constitute problems that need to be solved by appropriate electoral arrangements.

Another approach maintains that electoral systems, as well as other constitutional and institutional features, reflect the historical and cultural contexts to which they belong. It considers electoral system choices as the consequence of various kinds of diffusion. According to this perspective, the choice of political institutions in a society is the result of influence from other societies rather than a response to sociopolitical requirements. The rational and the historical/cultural approaches are not opposed to each other: an institutional choice may be both rational and imitated. Furthermore, the constitutional setting may determine the choice of particular institutions. This approach maintains that some institutional choices follow logically upon others, whereas some combinations tend to be inefficient and inappropriate. In a certain institutional setting, some electoral systems may be more attractive than others. The influence of other political institutions constitutes the third basic approach to the study of electoral system choice.

Consequently, the purpose of the study is to explore the influence of contextual factors on the choice of electoral systems for (direct) parliamentary elections during the post-war era. Because of the macroscopic approach, I shall not pay attention to the specific factors that led to the adoption of the electoral law in each country. Rather, my intention is to discover general patterns and create an explanatory framework of the adoption of electoral systems in the post- war era. Electoral system choice in democracies is separately analysed. A combination of induction and deduction is applied, since there is no established theory in this field of electoral system research. Associations between contextual factors and electoral system choice are analysed from three competing but simultaneously overlapping theoretical perspectives: a rational, a historical and cultural, and an institutional perspective.

Four variables are analysed within the rational perspective: ethnic/cultural diversity, country size, party system structure, and party system transformation. The former two are pure structural traits, whereas the latter two also include an actor- related dimension. To be sure, the book is concerned with contextual determinants but the role of political actors is not totally disregarded. More specifically, the study does not focus on actors but on the political context in which actors make their decisions on electoral provisions. This matter becomes particularly apparent when the relationship between party systems and electoral systems is analysed. Electoral system choice as a consequence of the party system structure is foremost an actor-related matter but has a structural dimension as well. As to the rational perspective, I consequently distinguish between structurally generated determinants and actor-related determinants.

The historical and cultural perspective consists of the following variables: colonial diffusion, regional diffusion and temporal diffusion. Although both colonial legacy and regional influence are concerned with imitating institutional choices of prestigious countries, I follow the common approach of treating colonial legacy as a distinct kind of diffusion. In addition to the hierarchical and spatial dimensions of imitating institutional choices embodied by colonial and regional diffusion, a temporal dimension is introduced by including a variable that regards electoral system choice as an epoch phenomenon. Within the institutional perspective, three variables are analysed: form of government, territorial organization and chamber structure. This perspective certainly rests upon a weaker theoretical basis than the other two; yet it is justified with respect to the interest of analysing constitution- building and the way countries tend to combine different institutions.

There are several reasons for choosing the year 1945 as starting point. All national elections prior to the twentieth century were held under either the plurality or the majority rule; thus, providing little variation of the dependent variable. Concerning electoral system choice during the first few decades of the twentieth century, the most essential feature was the change from majoritarian to proportional systems in several Western European countries. These changes are well documented in the literature and explanations have been provided. The rational perspective of this study is to a large extent based upon these findings. The frequency of newly independent countries increased rapidly as of 1945, having previously been quite low. Generally speaking, the end of the Second World War meant the beginning of a new era in practically every field. Taking these circumstances into consideration, it appears natural to focus on the postwar era, thereby following a rather common approach in macro-comparative political science.

Does electoral system choice matter?

Popular elections are the defining institution of modern democracy. The purpose of parliamentary elections is, first, to decide who will represent each constituency in the legislature, and, second, what the overall composition of the legislature by political party will be. By translating votes into seats, these decisions are managed by the particular electoral system used for each single election. Hence, the electoral method is a key factor in the political process: it largely determines who gets what, when and how. Electoral rules are commonly assumed to condition the chances of success of political parties and candidates in competitive elections.

Within the range of democratic institutions, several experts assert that the choice of electoral system is the most important one. Arend Lijphart (1994: 1), for instance, has described the electoral system as the most fundamental element of representative democracy. Furthermore, the electoral system is regarded by many as the principal key to reforming the political system. Electoral rules are easier to change than most other features of the political system, not least because the electoral law is quite rarely included in the constitution. Referring to Giovanni Sartori, the electoral system is ‘the most specific manipulative instrument of politics’ (1968: 273). To be sure, most of the theory of electoral systems presupposes a democratic form of government – however, the choice of electoral systems in non- democracies matters as well. For one thing, all countries were non- democratic when the first electoral provisions were introduced. Furthermore, the electoral method plays an important role in consolidating a democratic system. The institutional framework of a non- democratic nation may have a decisive impact on the development towards democracy.

As mentioned, the bulk of the electoral system research has dealt with political consequences of elections conducted under different electoral rules, while much less attention has been given to the understanding of cross- national variation in electoral systems. In a critical survey of electoral system research in the mid- 1980s, Lijphart (1985b) maintained that there was a shortage of studies of electoral system choice – in a volume published two decades later, Matthew S. Shugart (2005: 51) points out that this is still the case today, although some progress has been made. The purpose of writing this book is to fill part of this gap in the field of electoral system research. A great deal of the scientific discussion on the advantages and disadvantages of electoral systems has been concerned with the question of which electoral system should be chosen in each given context. It emphasizes the political outcomes of electoral system design as the principal criterion for choosing between different systems. If we reformulate the question and ask which system is likely to be chosen, we recognize that electoral systems are chosen in the context of existing institutional and cultural settings. The contextual framework shapes the preferences of political actors as to which electoral system is to be chosen, and constrains their negotiating strategies over alternative rule configurations. Obviously, both actors and structures matter in the process of choosing electoral systems. Explanations based on strategic calculations of political actors have been provided by, for instance, Carles Boix (1999), Josep M. Colomer (2004a) and Lijphart (1992a). The present study, in contrast, aims at explaining electoral system choice from a contextual perspective.

2 Research design

2.1 Methodological considerations

Basically, as the title suggests, this is a comparative study. Explanations of electoral system choice are sought for by comparing countries with different electoral systems. More precisely, associations between variables, and the mutual strength of explanatory variables, are studied by means of statistical analysis. There are two fundamental strategies of arriving at law-like generalizations about the relationship between two or several social phenomena: deduction and induction. Deduction means that we depart from a theoretical framework in the research field. On the basis of the theory, we derive hypotheses which are tested and either confirmed or rejected. As a result, the theory is developed, rejected or modified. When applying the inductive strategy, in contrast, we aim at detecting causal mechanisms by observing the empirical reality. The research process begins with charting the variation of the dependent variable, whereas the deductive strategy starts with selecting cases that show a variation in the explanatory variables.

When choosing between a deductive and an inductive approach, the amount of previous research in the area is decisive. If a comprehensive theory already exists, the deductive strategy should be applied. The inductive approach is preferable when such a theory is yet to be formulated. However, empirical studies are rarely totally deductive or inductive. In case of the latter, there is always an element of deduction involved. More exactly, although we lack a well-founded theoretical framework, we nevertheless have an idea of which settings to look into in order to find potential determinants of the phenomenon that we are studying. Accordingly, in the lack of a comprehensive theory of electoral system choice, this is primarily an inductive study. However, the independent variables are derived by using deduction. To elucidate, in the search of relevant variables, the present study makes use of previous attempts at, theoretically or empirically, explaining electoral system choice. The independent variables emanate from three broad settings: a rational, a historical and cultural, and an institutional context.

Two different operationalizations of the dependent variable are constructed in the following chapter: a detailed categorization of 15 electoral formulas and a broad typology of four categories of electoral systems (plurality, majority, proportional and mixed systems). The latter is subject to statistical analysis. With regard to proportionality, the typology resembles an ordinal classification: mixed systems are more proportional than plurality/majority systems, and proportional systems are more proportional than mixed systems. In a strict sense, however, the classification is not ordinal but categorical. For one thing, every plurality/majority system does not produce disproportional results, and all proportional systems do not result in proportional seat allocation. Furthermore, the inherent proportionality of PR systems and non-proportional nature of majoritarian systems should be regarded as effects rather than characteristics of electoral systems. As a consequence, despite the common classification of electoral systems approximate to a scale of different degrees of proportionality, the four-scale classification is nonetheless categorical. Furthermore, definitions of the categories do not meet the criteria for a pure ordinal scale. Logistic regression is therefore preferred to OLS (ordinary least squares) regression. Since the dependent variable consists of four values, multinomial logistic regression is primarily applied. The multiple analyses are preceded by descriptive and bivariate analyses in which simpler statistical methods are used. Due to the different nature of the explanatory variables, the classification and identification of electoral systems varies from case to case; some independent variables require a more detailed categorization, whereas the four-scale classification constitutes the logical basis of comparison in other contexts.

Data on electoral systems and the explanatory variables is collected from a variety of sources. Regarding the dependent variable, the main sources are Elections in Africa (Nohlen et al. 1999), both volumes of Elections in Asia and the Pacific (Nohlen et al. 2001a; 2001b), Handbuch der Wahldaten Lateinamerikas und der Karibik (Nohlen 1993b), ‘Mixed Electoral Systems: A Conceptual and Empirical Survey’ (Massicotte and Blais 1999) and different volumes in the series Chronicle of Parliamentary Elections, provided by the Inter-Parliamentary Union, as well as the organization’s website, Parline Database on National Parliaments.

In addition to the sources above, data on the independent variables has been gathered from The International Almanac of Electoral History (Mackie and Rose 1974), Britannica Book of the Year (1999), Svensk uppslagsbok (different editions 1947–1955), The World in Figures (editions 1981 and 1987), The World Factbook (CIA website), Statsskick: att bygga demokrati (Karvonen 2003) and Political Systems of the World (Derbyshire and Derbyshire 1989; 1999).

2.2 The research population

The unit of analysis is electoral system choice. All choices of electoral systems in independent countries between 1945 and 2007 are observed, irrespective of the way in which they have been adopted. This comprises the first choice of electoral systems in countries that have declared independence during the relevant period as well as every major electoral system change in countries that previously have had electoral laws as independent nations. There are, however, several countries in which electoral provisions existed and elections had been held before independence. A common feature of these countries is that they have been under some kind of foreign rule. It means that the electoral system had already been chosen when the country became independent. The study consists of independent countries only – autonomous territories that may become independent in the future are not paid attention to.

The apparent question is whether I should observe the year of electoral system choice for each country irrespective of whether independence had been attained at that point in time or not? Considering that there are several autonomous regions with electoral institutions in the world, the answer is no. Several of these territories may become independent in the future but they are, nonetheless, excluded from the study since it deals, as mentioned, exclusively with independent countries. However, an electoral system choice that was made a few years before independence is not less important than a choice that was made shortly after independence was proclaimed. Therefore, electoral system choices prior to independence cannot be disregarded.

The dilemma is to decide the timing of electoral system choice in countries that already had electoral provisions when independence was declared. It seems reasonable to consider the year of independence as the year of electoral system choice. In most cases, the introduction of pre-independence electoral provisions formed a part of the independence process. However, there are some exceptions to this rule in the Caribbean and the Pacific. Antigua and Barbuda, St Lucia, and St Kitts and Nevis, for example, held elections several decades before finally receiving independence.

The first general elections in Antigua and Barbuda took place in 1951; that is, 30 years before the archipelago became independent (Hillebrands 1993: 22).¹ Yet, the coded year of electoral system choice in this study is 1981. If 1951 is considered as the official year of choosing an electoral system, we may ask why other non-independent territories in which general elections have been held are excluded from the analysis. The obvious answer is, as already stated, that the study deals with electoral systems in independent countries. When it comes down to it, the country in question had the opportunity to reject the old electoral system and adopt another when independence was achieved. As for countries without pre-independence electoral institutions, the timing of the relevant electoral laws determines the coding of electoral system choice. If information on the date of the electoral law is missing, the year of the first parliamentary elections under the new electoral system is decisive. Another reason to consider the timing of independence as the year of electoral system choice is that information on the first pre-independence elections is often inadequate. Besides, if the first elections were decisive, only cases of electoral system change would fulfil the criteria of democracy stipulated below.

Election results in democratic countries are usually regarded as more important than electoral outcomes in non-democratic countries. Therefore, electoral system choice is of greater scientific interest in countries that allow multi-party competition. If all ca...

Table of contents

• Cover
• Title
• Copyright
• Dedication
• Contents
• List of illustrations
• Preface
• Acknowedgements
• 1 Introduction
• 2 Research design
• 3 Electoral systems
• 4 The rational perspective
• 5 History and culture: patterns of diffusion
• 6 The institutional perspective
• 7 Multivariate patterns
• 8 Conclusion
• Appendix
• Notes
• Bibliography
• Index